Face Generation

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data

You'll be using two datasets in this project:

  • MNIST
  • CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

In [1]:
#data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)
Found mnist Data
Found celeba Data

Explore the Data

MNIST

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

In [2]:
show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')
Out[2]:
<matplotlib.image.AxesImage at 0x7f18867accc0>

CelebA

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

In [3]:
show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))
Out[3]:
<matplotlib.image.AxesImage at 0x7f18866a4d68>

Preprocess the Data

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network

You'll build the components necessary to build a GANs by implementing the following functions below:

  • model_inputs
  • discriminator
  • generator
  • model_loss
  • model_opt
  • train

Check the Version of TensorFlow and Access to GPU

This will check to make sure you have the correct version of TensorFlow and access to a GPU

In [4]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf
import pickle as pkl

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))
TensorFlow Version: 1.0.0
Default GPU Device: /gpu:0

Input

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

  • Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
  • Z input placeholder with rank 2 using z_dim.
  • Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

In [5]:
import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    input_real = tf.placeholder(tf.float32,shape=[None,image_width,image_height,image_channels],name="input_real")
    input_z    = tf.placeholder(tf.float32,shape=[None,z_dim],name="input_z")
    lr         = tf.placeholder(tf.float32,name = "learning_rate")

    return input_real,input_z,lr


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)
Tests Passed

Discriminator

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the generator, tensor logits of the generator).

In [6]:
def discriminator(images, reuse=False,alpha = 0.2):
    """
    Create the discriminator network
    :param image: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    
    
    with tf.variable_scope('discriminator', reuse=reuse):
        
        # Input layer is 28x28x channel
        # now it is 14x14x64
        conv1 = tf.layers.conv2d(images, 64, 5, strides=2, padding='same')
        conv1 = tf.maximum(alpha * conv1, conv1)
    
        
        # now it is 14x14x128
        conv2 = tf.layers.conv2d(conv1, 128, 5, strides=2, padding='same',use_bias=False)
        conv2 = tf.layers.batch_normalization(conv2, training=True)
        conv2 = tf.maximum(alpha * conv2, conv2)
 
        # now it is 7*7*256
        conv3 = tf.layers.conv2d(conv2, 256, 5, strides=2, padding='same',use_bias=False)
        conv3 = tf.layers.batch_normalization(conv3, training=True)
        conv3 = tf.maximum(alpha * conv3, conv3)

        # Flatten it
        fc1 = tf.contrib.layers.flatten(conv3)
        logits = tf.layers.dense(fc1, 1)
        out = tf.sigmoid(logits)
        
        return out, logits



"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)
Tests Passed

Generator

Implement generator to generate an image using z. This function should be able to reuse the variabes in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

In [7]:
def generator(z, out_channel_dim, is_train=True,alpha=0.2):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    with tf.variable_scope('generator', reuse= not is_train):

        # First fully connected layer
        fc1 = tf.layers.dense(z, 7*7*256,use_bias=False)
        # Reshape it to start the convolutional stack
        # 7x7x256 now
        fc1 = tf.reshape(fc1, (-1, 7, 7, 256))
        fc1 = tf.layers.batch_normalization(fc1, training=is_train)
        fc1 = tf.maximum(alpha * fc1, fc1)
        
        # 14x14x128 now
        conv1 = tf.layers.conv2d_transpose(fc1, 128, 5, strides=2, padding='same',use_bias=False)
        conv1 = tf.layers.batch_normalization(conv1, training=is_train)
        conv1 = tf.maximum(alpha * conv1, conv1)
        
        # 28x28x64 now
        conv2 = tf.layers.conv2d_transpose(conv1, 64, 5, strides=1, padding='same',use_bias=False)
        conv2 = tf.layers.batch_normalization(conv2, training=is_train)
        conv2 = tf.maximum(alpha * conv2, conv2)
    
        
        # Output layer 28x28xout_channel_dim
        logits = tf.layers.conv2d_transpose(conv2, out_channel_dim, 3, strides=2, padding='same')
        
        out = tf.tanh(logits)
        
        return out
"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)
Tests Passed

Loss

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

  • discriminator(images, reuse=False)
  • generator(z, out_channel_dim, is_train=True)
In [8]:
def model_loss(input_real, input_z, out_channel_dim,smooth=0.1):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    
    # Build the model
    g_model = generator(input_z, out_channel_dim,is_train=True)
    # g_model is the generator output

    # get the 
    d_model_real, d_logits_real = discriminator(input_real,reuse=False)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)
    
    d_loss_real = tf.reduce_mean(
                  tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, 
                                                          labels=tf.ones_like(d_logits_real) * (1 - smooth)))
    d_loss_fake = tf.reduce_mean(
                  tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, 
                                                          labels=tf.zeros_like(d_logits_fake))) 
    d_loss = d_loss_real + d_loss_fake

    g_loss = tf.reduce_mean(
             tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake,
                                                     labels=tf.ones_like(d_logits_fake)))
    
    return d_loss, g_loss


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)
Tests Passed

Optimization

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

In [9]:
def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # Get the trainable_variables, split into G and D parts
    t_vars = tf.trainable_variables()
    g_vars = [var for var in t_vars if var.name.startswith('generator')]
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]

    # Tell TensorFlow to update the population statistics while training
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_train_opt = tf.train.AdamOptimizer(learning_rate,beta1=beta1).minimize(d_loss, var_list=d_vars)
   
    # Tell TensorFlow to update the population statistics while training
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        g_train_opt = tf.train.AdamOptimizer(learning_rate,beta1=beta1).minimize(g_loss, var_list=g_vars)
    
    return d_train_opt, g_train_opt


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)
Tests Passed

Neural Network Training

Show Output

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

In [10]:
"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train

Implement train to build and train the GANs. Use the following functions you implemented:

  • model_inputs(image_width, image_height, image_channels, z_dim)
  • model_loss(input_real, input_z, out_channel_dim)
  • model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

In [11]:
!mkdir checkpoints
In [12]:
def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # TODO: Build Model
    
    image_width = data_shape[1]
    image_height = data_shape[2]
    image_channels = data_shape[3]

    input_real,input_z,lr = model_inputs(image_width,image_height,image_channels,z_dim)
    
   
    d_loss,g_loss = model_loss(input_real,input_z,image_channels)
    d_train_opt, g_train_opt = model_opt(d_loss,g_loss,lr,beta1)
    
    # record train steps
    steps = 0
    eps = 0
    saver = tf.train.Saver()
    # record train losses
    losses = []
    with tf.Session() as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                # Sample random noise for G
                batch_z = np.random.uniform(-1, 1, size=(batch_size, z_dim))
            
                # Run optimizers
                
                _ = sess.run(d_train_opt, feed_dict={input_real: batch_images,
                                                     input_z: batch_z,
                                                     lr:learning_rate})
                _ = sess.run(g_train_opt, feed_dict={input_real: batch_images,
                                                     input_z: batch_z,
                                                     lr:learning_rate})
                
                steps +=1
                
                if steps % 100 == 0:
                    train_loss_d = sess.run(d_loss, {input_z: batch_z, input_real: batch_images})
                    train_loss_g = g_loss.eval({input_z: batch_z})
                    # Save losses to view after training
                    losses.append((train_loss_d, train_loss_g))
                    
                    print("Epoch {} / Batch {}...".format(eps,steps),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g)) 
                    show_generator_output(sess,25,input_z,image_channels,data_image_mode)
                    
           
            # At the end of each epoch, get the losses and print them out
            train_loss_d = sess.run(d_loss, {input_z: batch_z, input_real: batch_images})
            train_loss_g = g_loss.eval({input_z: batch_z})
            
            print("Epoch {} / Batch {}...".format(eps,steps),
                  "Discriminator Loss: {:.4f}...".format(train_loss_d),
                  "Generator Loss: {:.4f}".format(train_loss_g))  
            
            show_generator_output(sess,4,input_z,image_channels,data_image_mode)
            eps +=1
        
        #saver.save(sess, './checkpoints/gan.ckpt')
    
    with open('losses.pkl', 'wb') as f:
        pkl.dump(losses, f)
                
                

MNIST

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

In [13]:
batch_size = 32
z_dim = 128
learning_rate = 0.0002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)
Epoch 0 / Batch 100... Discriminator Loss: 1.2841... Generator Loss: 3.0431
Epoch 0 / Batch 200... Discriminator Loss: 0.6906... Generator Loss: 1.6483
Epoch 0 / Batch 300... Discriminator Loss: 0.6427... Generator Loss: 1.4824
Epoch 0 / Batch 400... Discriminator Loss: 0.7952... Generator Loss: 1.1218
Epoch 0 / Batch 500... Discriminator Loss: 0.5357... Generator Loss: 2.0984
Epoch 0 / Batch 600... Discriminator Loss: 0.5034... Generator Loss: 2.2360
Epoch 0 / Batch 700... Discriminator Loss: 0.4519... Generator Loss: 2.5068
Epoch 0 / Batch 800... Discriminator Loss: 0.6760... Generator Loss: 1.3813
Epoch 0 / Batch 900... Discriminator Loss: 0.4562... Generator Loss: 2.2893
Epoch 0 / Batch 1000... Discriminator Loss: 0.4521... Generator Loss: 2.3836
Epoch 0 / Batch 1100... Discriminator Loss: 0.7442... Generator Loss: 1.1797
Epoch 0 / Batch 1200... Discriminator Loss: 0.5102... Generator Loss: 2.0838
Epoch 0 / Batch 1300... Discriminator Loss: 0.4594... Generator Loss: 2.4241
Epoch 0 / Batch 1400... Discriminator Loss: 0.5820... Generator Loss: 1.7182
Epoch 0 / Batch 1500... Discriminator Loss: 0.4962... Generator Loss: 2.0365
Epoch 0 / Batch 1600... Discriminator Loss: 0.7802... Generator Loss: 1.1384
Epoch 0 / Batch 1700... Discriminator Loss: 0.4610... Generator Loss: 2.2019
Epoch 0 / Batch 1800... Discriminator Loss: 0.5246... Generator Loss: 1.9280
Epoch 0 / Batch 1875... Discriminator Loss: 0.7515... Generator Loss: 1.2514
Epoch 1 / Batch 1900... Discriminator Loss: 0.4902... Generator Loss: 2.0858
Epoch 1 / Batch 2000... Discriminator Loss: 0.7496... Generator Loss: 1.2534
Epoch 1 / Batch 2100... Discriminator Loss: 0.6814... Generator Loss: 1.3355
Epoch 1 / Batch 2200... Discriminator Loss: 0.5298... Generator Loss: 1.8467
Epoch 1 / Batch 2300... Discriminator Loss: 0.8189... Generator Loss: 1.0703
Epoch 1 / Batch 2400... Discriminator Loss: 0.4763... Generator Loss: 2.5164
Epoch 1 / Batch 2500... Discriminator Loss: 0.4210... Generator Loss: 2.5593
Epoch 1 / Batch 2600... Discriminator Loss: 0.4541... Generator Loss: 2.3474
Epoch 1 / Batch 2700... Discriminator Loss: 0.5421... Generator Loss: 1.8613
Epoch 1 / Batch 2800... Discriminator Loss: 0.5663... Generator Loss: 1.9987
Epoch 1 / Batch 2900... Discriminator Loss: 0.5191... Generator Loss: 1.9468
Epoch 1 / Batch 3000... Discriminator Loss: 0.4438... Generator Loss: 2.5246
Epoch 1 / Batch 3100... Discriminator Loss: 0.5251... Generator Loss: 1.8881
Epoch 1 / Batch 3200... Discriminator Loss: 0.6009... Generator Loss: 1.5665
Epoch 1 / Batch 3300... Discriminator Loss: 0.5857... Generator Loss: 2.2564
Epoch 1 / Batch 3400... Discriminator Loss: 1.1616... Generator Loss: 4.3771
Epoch 1 / Batch 3500... Discriminator Loss: 0.4355... Generator Loss: 2.4941
Epoch 1 / Batch 3600... Discriminator Loss: 0.5060... Generator Loss: 2.4862
Epoch 1 / Batch 3700... Discriminator Loss: 0.5425... Generator Loss: 2.0955
Epoch 1 / Batch 3750... Discriminator Loss: 1.1665... Generator Loss: 2.2386
In [14]:
with open('losses.pkl', 'rb') as f:
    losses= pkl.load(f)
    losses = np.array(losses)
    fig, ax = pyplot.subplots()
    pyplot.plot(losses.T[0], label='Discriminator', alpha=0.5)
    pyplot.plot(losses.T[1], label='Generator', alpha=0.5)
    pyplot.title("Training Losses")
    pyplot.legend()

CelebA

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

In [15]:
batch_size = 32
z_dim = 128
learning_rate = 0.0002
beta1 = 0.5


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)
Epoch 0 / Batch 100... Discriminator Loss: 0.5225... Generator Loss: 2.5144
Epoch 0 / Batch 200... Discriminator Loss: 0.6946... Generator Loss: 2.8764
Epoch 0 / Batch 300... Discriminator Loss: 1.2523... Generator Loss: 0.8005
Epoch 0 / Batch 400... Discriminator Loss: 1.1450... Generator Loss: 0.6978
Epoch 0 / Batch 500... Discriminator Loss: 0.9455... Generator Loss: 1.1040
Epoch 0 / Batch 600... Discriminator Loss: 0.5556... Generator Loss: 1.9163
Epoch 0 / Batch 700... Discriminator Loss: 0.5919... Generator Loss: 2.9024
Epoch 0 / Batch 800... Discriminator Loss: 0.4131... Generator Loss: 3.5333
Epoch 0 / Batch 900... Discriminator Loss: 0.3939... Generator Loss: 3.3411
Epoch 0 / Batch 1000... Discriminator Loss: 0.5200... Generator Loss: 6.3571
Epoch 0 / Batch 1100... Discriminator Loss: 0.3349... Generator Loss: 8.4472
Epoch 0 / Batch 1200... Discriminator Loss: 0.3450... Generator Loss: 6.3671
Epoch 0 / Batch 1300... Discriminator Loss: 1.1551... Generator Loss: 0.6625
Epoch 0 / Batch 1400... Discriminator Loss: 0.6051... Generator Loss: 2.3302
Epoch 0 / Batch 1500... Discriminator Loss: 0.3630... Generator Loss: 10.4310
Epoch 0 / Batch 1600... Discriminator Loss: 0.3460... Generator Loss: 5.4350
Epoch 0 / Batch 1700... Discriminator Loss: 0.3539... Generator Loss: 4.5012
Epoch 0 / Batch 1800... Discriminator Loss: 0.4691... Generator Loss: 2.1809
Epoch 0 / Batch 1900... Discriminator Loss: 0.3733... Generator Loss: 3.6423
Epoch 0 / Batch 2000... Discriminator Loss: 0.3621... Generator Loss: 9.0185
Epoch 0 / Batch 2100... Discriminator Loss: 0.3357... Generator Loss: 5.3223
Epoch 0 / Batch 2200... Discriminator Loss: 3.7507... Generator Loss: 0.0581
Epoch 0 / Batch 2300... Discriminator Loss: 0.3567... Generator Loss: 4.8278
Epoch 0 / Batch 2400... Discriminator Loss: 0.3313... Generator Loss: 6.7689
Epoch 0 / Batch 2500... Discriminator Loss: 0.3307... Generator Loss: 13.5750
Epoch 0 / Batch 2600... Discriminator Loss: 0.9854... Generator Loss: 0.9767
Epoch 0 / Batch 2700... Discriminator Loss: 0.3349... Generator Loss: 9.6795
Epoch 0 / Batch 2800... Discriminator Loss: 0.3412... Generator Loss: 8.9721
Epoch 0 / Batch 2900... Discriminator Loss: 0.3282... Generator Loss: 11.1475
Epoch 0 / Batch 3000... Discriminator Loss: 5.7631... Generator Loss: 7.5518
Epoch 0 / Batch 3100... Discriminator Loss: 0.3315... Generator Loss: 13.3830
Epoch 0 / Batch 3200... Discriminator Loss: 0.3372... Generator Loss: 10.8214
Epoch 0 / Batch 3300... Discriminator Loss: 0.3307... Generator Loss: 6.3121
Epoch 0 / Batch 3400... Discriminator Loss: 0.3299... Generator Loss: 8.0404
Epoch 0 / Batch 3500... Discriminator Loss: 0.3317... Generator Loss: 14.4564
Epoch 0 / Batch 3600... Discriminator Loss: 0.3438... Generator Loss: 4.2826
Epoch 0 / Batch 3700... Discriminator Loss: 0.3320... Generator Loss: 8.7700
Epoch 0 / Batch 3800... Discriminator Loss: 0.3569... Generator Loss: 3.9665
Epoch 0 / Batch 3900... Discriminator Loss: 0.3838... Generator Loss: 3.5584
Epoch 0 / Batch 4000... Discriminator Loss: 0.7474... Generator Loss: 2.7833
Epoch 0 / Batch 4100... Discriminator Loss: 0.3301... Generator Loss: 11.2567
Epoch 0 / Batch 4200... Discriminator Loss: 0.3323... Generator Loss: 10.1797
Epoch 0 / Batch 4300... Discriminator Loss: 0.3273... Generator Loss: 12.6743
Epoch 0 / Batch 4400... Discriminator Loss: 0.3449... Generator Loss: 12.6680
Epoch 0 / Batch 4500... Discriminator Loss: 0.3327... Generator Loss: 12.8955
Epoch 0 / Batch 4600... Discriminator Loss: 0.3492... Generator Loss: 4.8613
Epoch 0 / Batch 4700... Discriminator Loss: 0.3675... Generator Loss: 3.4819
Epoch 0 / Batch 4800... Discriminator Loss: 0.3288... Generator Loss: 8.2498
Epoch 0 / Batch 4900... Discriminator Loss: 0.3304... Generator Loss: 6.9006
Epoch 0 / Batch 5000... Discriminator Loss: 0.8386... Generator Loss: 1.2381
Epoch 0 / Batch 5100... Discriminator Loss: 0.3400... Generator Loss: 8.7053
Epoch 0 / Batch 5200... Discriminator Loss: 0.6368... Generator Loss: 1.5577
Epoch 0 / Batch 5300... Discriminator Loss: 0.3304... Generator Loss: 8.0844
Epoch 0 / Batch 5400... Discriminator Loss: 0.3359... Generator Loss: 5.0490
Epoch 0 / Batch 5500... Discriminator Loss: 0.3288... Generator Loss: 6.7634
Epoch 0 / Batch 5600... Discriminator Loss: 0.3309... Generator Loss: 10.5659
Epoch 0 / Batch 5700... Discriminator Loss: 0.3296... Generator Loss: 7.9620
Epoch 0 / Batch 5800... Discriminator Loss: 0.3299... Generator Loss: 7.3849
Epoch 0 / Batch 5900... Discriminator Loss: 0.3442... Generator Loss: 7.1332
Epoch 0 / Batch 6000... Discriminator Loss: 0.3288... Generator Loss: 8.0923
Epoch 0 / Batch 6100... Discriminator Loss: 0.3305... Generator Loss: 6.6738
Epoch 0 / Batch 6200... Discriminator Loss: 0.3270... Generator Loss: 8.9220
Epoch 0 / Batch 6300... Discriminator Loss: 0.3289... Generator Loss: 6.3978
Epoch 0 / Batch 6331... Discriminator Loss: 0.3290... Generator Loss: 7.3942
In [16]:
with open('losses.pkl', 'rb') as f:
    losses= pkl.load(f)
    losses = np.array(losses)
    fig, ax = pyplot.subplots()
    pyplot.plot(losses.T[0], label='Discriminator', alpha=0.5)
    pyplot.plot(losses.T[1], label='Generator', alpha=0.5)
    pyplot.title("Training Losses")
    pyplot.legend()

Submitting This Project

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.